1

Detection of Abnormal Wear Particles in Hydraulic Fluids via Electromagnetic Sensor and Particle Imaging

Technologies

Thomas Barraclough and Patrick Henning, Spectro Scientific

Andrew Velasquez, Jordan Saikia and Paul Michael, Milwaukee School of Engineering

Abstract: The importance of particle analysis in hydraulic fluid power systems has long been established. With the

increased use of high-pressure pumps and precision electrohydraulic motion control systems, the characterization

of metallic wear debris is gaining importance. An accurate morphological analysis of wear particles provides for a

complete characterization of the wear mechanisms and the severity of wear conditions. In this investigation,

particle imaging, electromagnetic, ferrographic, and spectroscopic methods were used to evaluate hydraulic oil

samples collected from machines known to be generating abnormal wear particles. Analytical ferrography

confirmed the presence of large reworked ferrous particles and small rubbing wear particles in the oil samples. ICP

emission spectroscopy was limited in its ability to detect both small and large ferrous particles. An automated

particle imaging system incorporating dual electromagnetic sensors counted the number of ferrous wear particles

larger than 25µm in the oil and measured the total ferrous concentration in parts per million. The ratio of large to

small ferrous wear particles and their concentration revealed the severity of wear occurring in the machines. These

results demonstrate the diagnostic advantage of combining magnetic and particle imaging sensors in an integrated

oil analysis system.

Keywords: wear, ferrography, magnetic sensors, particle imaging, hydraulic fluids

1. Introduction

Hydraulic transmissions are used in agriculture, construction, manufacturing, mining and transportation

equipment. Hydraulic fluid power is the preferred technology in these applications because it can transmit high

forces with precision control. Maintenance of hydraulic fluids is required because wear particles can reduce the

efficiency and reliability of a hydraulic system. If appropriate measures are not taken, then catastrophic failure may

occur. Particulate contamination is believed to cause most of the machine faults in hydraulic systems. [1] When

abnormal wear occurs in a hydraulic system, relatively large particles are created. These particles cause abrasive

wear of pump port plates, scoring in cylinder bores, and damage hydraulic valves. The valve spool shown in Figure

1 is from the hydraulic pitch control system of a wind turbine. High levels of contamination in the hydraulic fluid

damaged the spool and altered the geometry of the metering grooves, leading to catastrophic over-speeding of

the machine.

Figure 1: Damaged surfaces of hydraulic spool valves from the pitch control system of a wind turbine exhibiting a) fatigue cracking, b) displaced metal in metering groove, c) surface abrasion and wear debris in metering groove

2

2. Background

2.1. Particle Count Analysis

Hydraulic equipment is uniquely sensitive to particulate contamination because the fluid is the power transmission

medium. To efficiently transmit power the pressure envelope, which stores and conveys power, must be isolated

from the low-pressure zones where fluid power is converted back into mechanical energy. This is accomplished by

minimizing the gaps between surfaces that are in relative motion such as at the side-plate/rotor interface in pumps

or between the spool and valve body in control valves. Typical clearances for pumps and valves utilized in hydraulic

systems are shown in Table 1. Gap heights are measured in “micrometers” (µm), which is one millionth of a

meter. The visibility limit of the human eye is approximately 40μm. Thus, the maximum clearance in hydraulic

components is below the limit of observation with the unaided eye. In order to prevent wear at critical gaps within

hydraulic equipment, the particulate contamination level of the fluid must be kept to a minimum [1].

Table 1: Typical gap heights in hydraulic pumps and valves (ISO 12669)

Automatic particle counters provide an accurate means of determining hydraulic fluid contamination levels. [2]

Fluid cleanliness levels are typically expressed in terms of particle contamination codes. These codes describe the

number of particles per unit volume within a specified size range. In commercial hydraulic applications, the ISO

4406 contamination code is used to quantify fluid contamination levels [3]. In the ISO 4406 classification system,

the number of particles per milliliter (ml) of fluid greater than or equal to (≥) 4μm, 6μm, and 14μm are classified

using the ranges described in . These particle size ranges were selected because they correspond to the critical

gap heights in hydraulic components. A fluid that contains 3000 particles ≥4μm, 1000 particles ≥6μm and 200

particles ≥14μm would classified ISO Code 19/17/15. Since the particle level doubles at each scale increment, an

ISO Code 20/18/16 fluid contains twice as many particles as an ISO Code 19/17/15 fluid.

The Required Cleanliness Level (RCL) of the fluid in a hydraulic machine depends upon operating pressure,

component contamination sensitivity, system life expectancy, and other application-specific requirements.

Systems that operate at high pressures and incorporate servo-electric control valves have stringent requirements

for fluid cleanliness. Systems that operate at low pressures and utilize manual or solenoid directional control

valves can tolerate higher levels of contamination. An objective method for determining the RCL of hydraulic fluids

is provided in ISO 12669 Hydraulic Fluid Power – Method for determining the required cleanliness level (RCL) for a

system. [4] This procedure assigns weighting factors for:

Component Gap height, µm

Vane pump

• Side plate/rotor 5 - 13

• Vane tip/cam 0.5 - 1

Gear pump

• Side Plate/rotor 0.5 - 5

• Housing/gear tip 0.5 - 5

Piston pump

• Piston/bore 5 - 40

• Valve plate/cylinder 0.5 – 5

Servo valve

• Spool to sleeve 1 - 4

• Flapper wall 18 - 63

Directional control valve

• Control piston 2.5 - 23

• Cone valve 13 - 40

3

• Working pressure and duty cycle

• Component contamination sensitivity

• System life expectancy

• Cost of replacement

• Downtime expense

• Risk of additional hazards

The sum of system weightings is used to select the RCL. Based upon ISO 12669 criteria, the RCL is ISO 15/13/9 for

a large hydraulic excavator operating at high-intensity in a quarry where component contamination sensitivity and

prohibitive downtime costs are a concern. ISO 15/13/9 is a stringent cleanliness requirement. To put this RCL into

perspective, a fluid that contains 2.8 ppm of ISO Medium test dust has a minimum ISO Code of 19/18/16. A fluid

that is ISO 15/13/9 equates to < 0.2 ppm of ISO Medium test dust, which is below the detection limit of many

routine oil analysis methods. While 15/13/9 corresponds to less than 0.2 ppm of contamination, an ISO 15/13/9

fluid contains more than 160 particles ≥4µm, 80 particles ≥6µm, and 2.5 particles ≥4µm per milliliter as shown in

Table 2.

Table 2: Allocation of ISO 4406 scale numbers (particles/ml) [3]

Number of particles per milliliter Scale #

More than Up to and including Code

80,000 160,000 24

40,000 80,000 23

20,000 40,000 22

10,000 20,000 21

5,000 10,000 20

2,500 5,000 19

1,300 2,500 18

640 1,300 17

320 640 16

160 320 15

80 160 14

40 80 13

20 40 12

10 20 11

5 10 10

2.5 5 9

According to ISO 4406:2017, the minimum number of counts required to establish an ISO code is 20. When the

raw data in one of the size ranges results in a particle count of fewer than 20 particles, the scale number for that

size range is labelled with the greater than or equal to (≥) symbol. Detection of as few as 2.5 particles/ml may be

accomplished by analyzing 3 ml of fluid in triplicate. Hence, particle counting is uniquely capable of verifying that a

fluid meets the stringent cleanliness requirements of hydraulic fluid power systems.

2.2. Atomic Emission Spectroscopy

In terms of the total number of samples processed, atomic emission spectroscopy is perhaps the most widely used

oil analysis technique. Atomic emission spectroscopy is extensively employed because it permits simultaneous

measurement of ppm concentrations of wear metals, contaminants, and additives in oil samples. In Rotating Disc

4

Electrode atomic emission spectroscopy (RDE-AES), a large electric potential is set up between the rotating disc

and an electrode, with the oil sample in the gap between them. The discharged of the electric potential across this

gap creates a high temperature electric arc that vaporizes the sample forming a plasma. In Inductively Coupled

Plasma atomic emission spectroscopy (ICP-AES), an oil sample is aspirated through a high-temperature plasma of

ionized gas. In both methods, the exposure of the sample to extreme high temperatures within the plasma causes

atomic radiation emissions specific to each element. Since no two elements have the same pattern of spectral

lines, the elements can be differentiated via spectroscopy. The intensity of the emitted light is proportional to the

quantity of the element present in the sample, allowing the concentration of that element to be determined.

While AES techniques are readily adapted to automation, transport of ferrous particles > 6µm to the emission

source is inefficient. Thus, atomic emission spectroscopy is non-quantitative for oils that contain medium to large

ferrous wear particles. [5]

2.3. Analytical Ferrography

Analytical ferrography is another method that is used to test for the presence of wear debris in oil samples. In

analytical ferrography, a small amount of oil is dispersed in a hydrocarbon solvent and transferred to a thistle tube

where it is metered through a capillary onto the surface of a microscope slide. As the fluid passes across the

horizontally inclined surface of the slide, wear particles align with an adjacent magnetic field, depositing strings of

wear particles perpendicular to the direction of fluid flow. Large ferrous particles have a high magnetic

susceptibility and deposit at the inlet of the ferrogram. Smaller ferrous particles have a low magnetic

susceptibility and deposit at the outlet of the ferrogram. Nonferrous particles also accumulate on the slide,

precipitating with a random orientation. After the sample has passed across the magnetic gradient, the ferrogram

slide is washed with solvent and microscopically examined. Particle size, color, shape, and magnetic susceptibility

are characterized to in order provide a qualitative assessment of wear and the prevalent mechanisms of particle

generation. Although analytical ferrography is very effective at capturing large wear particles and can reveal the

fundamental wear mechanisms occurring in a lubrication system, it is a time-consuming test that is not well-suited

for automation. As a result, its application is often limited to high-value assets.

2.4. Magnetic Wear Particle Detection

Inline magnetic chip detectors have been used for helicopters gearbox and aircraft engine condition assessment

for over two decades. [6] These sensors function by monitoring the disturbance to a magnetic field caused by the

passing of a metallic particle through a sensing coil. The particle couples with the sensing field to varying degrees

as it passes the coil, producing a characteristic output signal. The amplitude and phase of the output is used to

identify the size and composition (magnetic vs non-magnetic) of the particle. While inline magnetic chip detectors

can perform well certain aviation applications, they are not well suited for use in hydraulic systems because inline

systems cannot detect particles that are smaller than 175µm. [7, 8]

In the present study, oil samples were collected from hydraulic machines known to be generating abnormal wear

particles and evaluated using particle counting, atomic emission, ferrographic, and magnetic particle detection

methods. A comparison of the results is used to discuss the ability of each method to distinguish between normal

and abnormal wear particle generation in fluid power systems.

3. Methods and Materials

3.1. Instrumentation

Particle counts were collected using the Spectro Scientific’s LaserNet 230 particle counter and classification

system. [9] This direct particle imaging system employs an Infrared laser, a flow cell, and a high-speed charge-

coupled device (CCD camera) to capture particle images. When a particle passes through the flow cell, the infrared

beam illuminating the flow cell casts a shadow of the particle that is projected onto the CCD chip. Through analysis

of pixilated shapes, the optical system measures the size distribution of particles from ≥4 to 100 µm. Particle

counts in the ≥4µm, ≥6µm, and ≥14µm size ranges are used to determine the ISO contamination code. Particles

5

that are larger than 20µm are classified based upon their pixelated images using an algorithm inspired by biological

neural networks. This artificial intelligence system recognizes particle shapes and classifies them as cutting wear

particles, fatigue wear particles, sliding wear particles, water droplets, air bubbles, fibers, or nonmetallic

contaminants. Tests were conducted in accordance with the ASTM D7596 – Standard Test Method for Automatic

Particle Counting and Particle Shape Classification of Oils [10].

In addition to a particle imaging system, the LaserNet 230 is equipped with twin magnetometers for the detection

of ferrous particles. These sensors measure a change in the inductance of a coil when ferrous particles traverse an

electromagnetic field. The first, smaller magnetometer counts the number of ferrous particles that are >25µm.

The second, larger magnetometer measures the total concentration of iron in ppm. It is important to note that

the measurements produced by the two magnetometers do not necessarily correlate because they quantify

different aspects of the ferrous wear particle distribution. The large magnetometer measures the concentration of

“normal” rubbing wear particles, while the small magnetometer counts the number of “abnormal” wear particles

in the fluid. According to ASTM D7684, normal rubbing wear particles are free metal platelets with smooth

surfaces, from approximately 0.5 to 15µm in major dimension. [11] The major dimension-to-thickness ratios of

normal rubbing wear particles range from 10:1 for larger particles to 3:1 for smaller. Ferrous particles larger than

15µm are generally produced by high-energy surface interactions that occur at the onset of severe wear.

According to ASTM D7684, severe wear particles are free metal particles >15µm with a major dimension-to-

thickness ratio between 5:1 and 30:1. From a practical standpoint, the cut-off for the minimum particle size

corresponding to abnormal ferrous wear particles depends upon the specific triboelements that comprise the

sliding, rolling, or abrasive contacts within a machine. In this investigation, the cut-off for abnormal wear was

25µm in equivalent spherical diameter. Particles above this size range were tallied by the small magnetometer

and classified by the optical imaging system.

Several oil analysis techniques were used to evaluate the ability of the twin magnetometer/optical imaging system

to distinguish between normal and abnormal wear. Wear metal analysis was conducted using a Perkin Elmer

Optima 5300 DV Inductively Coupled Plasma – Optical Emission Spectrometer. The spectrometer was calibrated

using multi-element oil-based calibration standards. Samples were diluted 10:1 by volume in petroleum solvent

and evaluated after testing with a solvent blank to ensure no sample-to-sample carryover. The concentration of

20 elements, including iron were reported in ppm. The concentration of iron was also measured using a Spectro

Scientific Ferro Check 2000 portable magnetometer. The Ferro Check 2000 works by sensing disruption of a

magnetic field that is generated due to the presence of ferrous debris in the oil. Operation involves simply drawing

the sample, placing it in the instrument and using the touchscreen to complete the analysis and view the results.

The FerroCheck magnetometer can detect particles from nanometers to millimeters in size and has a sensitivity

range of 0-2500 ppm with a limit of detection of less than 5 ppm. Qualitative analysis of the ferrous particle

concentration was performed using a Spectro Scientific T2FM Ferrogram Maker. In order to prepare ferrograms, 3

ml of hydraulic fluid was combined with 1 ml of heptane and transferred to a thistle tube where it was metered

through a capillary onto the ferrogram slide. Microscopic examination of the ferrogram was performed and the

size, shape, and composition of the particles was assessed using the ASTM D7690 – Standard Practice for

Microscopic Characterization of Particles from In-Service Lubricants by Analytical Ferrography [12].

3.2. Test Materials

Antiwear hydraulic fluid samples of ISO Viscosity Grade 46 were collected in clean sample bottles from thirty-six

individual hydraulic components that contained high levels of abnormal wear particles. A malfunctioning robotic

assembly system was identified as the root cause of high wear rates during roll-off testing. The components were

removed from inventory prior to shipment. Oil samples were drained from the machine case. Ferrous wear debris

was visible at the bottom of the sample containers upon settling.

6

4. Results

4.1. Reference Samples

The LaserNet 230 optical imaging system was factory calibrated using NIST traceable reference material SRM 2806.

[13] The calibration of the particle counter was validated using suspensions of ISO 12103-1, A3 Medium Test Dust

in MIL-PRF-5606 aviation grade hydraulic fluid. A concentrate suspension of ISO Medium test dust was prepared in

MIL-PRF-5606 fluid and aliquots of the oil/test dust mixture were weighed into clean sample bottles. Additional

diluent was added by weight, the samples were homogenized, and particle counts were performed.

Likewise, the magnetometers were factory calibrated using proprietary iron solutions. The calibration of the

magnetometers was validated using suspensions of carbonyl iron. Carbonyl iron is a highly pure iron prepared by

chemical decomposition of iron pentacarbonyl. [14] Three nominal sizes of carbonyl iron particles were used in

this study; 6µm, 12µm, and 24µm. A concentrate suspension of each grade was prepared in MIL-PRF-5606 fluid

and aliquots of the oil/carbonyl iron mixture were weighed into clean sample bottles. Additional diluent was

added by weight, the samples were homogenized, and the samples were evaluated. As shown in Figure 2, the

responses of the optical imaging and magnetometer systems were linear with respect to ISO medium test dust and

carbonyl iron concentrations. The 4 micron counts for the ISO Medium test dust were within one ISO code of the

theoretical value. The magnetometer measurements for total ferrous particle concentration were nearly twice as

high as the amount of carbonyl iron present in the reference samples. This was attributed to the fact that carbonyl

iron has a higher magnetic susceptibility than the ferrous alloys that were used to calibrate the sensors.

Figure 2: Verification of linear response in particle counter and magnetometer systems

4.2. Ferrography

Ferrograms were made of the carbonyl iron reference samples and the field samples. The ferrograms were

examined using an Olympus SZ51 stereo microscope. Images of the particles were captured with a high-resolution

camera and commercial imaging software. The particle distribution was qualitatively assessed. A comparison of

the ferrograms of carbonyl iron and field samples is shown in Figure 3. The ferrogram of carbonyl iron

predominantly consisted of very fine iron particles. The nominal size of the particles in is specimen was 6µm. Small

particles are less susceptible to magnetic forces. Consequently, the ferrous particles were distributed over a larger

portion of the ferrogram slide. The ferrogram of the field sample contains a mixture of large and small ferrous

particles with larger particles clustered toward the center of the glass substrate. The ASTM D7684 Standard Guide

for Microscopic Characterization of Particles from In-Service Lubricants defines size ranges for fine, small, medium

and large particles. Based upon D7684 criteria, the ferrograms of field samples exhibited a mixture of fine (<6µm),

7

small (6 to 14µm), medium (14 to 40µm) and large (40 to 100 µm) wear particles. [11] The morphologies of the

fine and small ferrous particles were typical of rubbing wear. The medium and large wear particles were thin and

flat platelets, with a major dimension ratio to thickness ratio of > 30:1. These particles appear to have been

reworked; that is, they originally were thicker particles and have been squeezed through rolling contacts.

Reworked particles are a concern because they are an indication of the presence of large particles, and their

passage through a rolling contact potentially may lead to subsurface cracking and bearing fatigue. [12]

Figure 3: Ferrograms of a) carbonyl iron and b) field sample.

4.3. Particle Count Analysis

Particle counts were collected on the field samples. The counts at ≥4µm, ≥6µm, and ≥14µm, and histograms of the

ISO codes are shown in Figure 4. The mean contamination level was ISO Code 24/22/17. The ISO Codes exhibited

a normal distribution about the mean value. The variance in the particle counts decreased with particle size, while

the standard deviation of the ISO Code increased. This is due to the fact that particle count ranges increase

exponentially in the ISO Code system. Clearly fluids of ISO Code 24/22/17 exceed the RCL of a typical hydraulic

system. ISO 24/22/17 is an actionable contamination level in a proactive oil analysis program. However, ISO Codes

do not reveal if the particles are externally ingressed contaminants or internally generated wear debris.

Figure 4: a) Particle counts and b) ISO Codes of field samples

8

4.4. Particle Imaging

As the optical imaging system counted the number of particles passing through the flow cell, projected shadows of

particles with an equivalent circular diameter greater than 20µm were automatically photographed and classified.

An example of particle silhouettes that were produced by the field samples is shown in Figure 5. Note that some

of the particles have gaps or missing pixels. Particles that permit the transmission of light are classified as non-

metallic by image recognition algorithms. Non-metallic particles are comprised of amorphous or crystalline

substances such as organic gels, plastic, sand, or glass. Non-metallic particles have sufficient atomic interactions to

maintain a solid or semi-solid structure, but they are not necessarily hard. Based upon LNF 230 particle

classification algorithms, an average of 25% of the large particles in the field samples were non-metallic while 75%

were metallic. The metallic particles consisted of cutting wear, fatigue and sliding wear particles, the majority of

which were sliding wear. Sliding wear is wear produced by relative motion in the tangential plane of contact

between two surfaces. The resulting wear produces particles that are >15µm and several times longer than they

are wide. Sliding wear particles may exhibit surface striations. Their major dimension-to-thickness ratio is

approximately 10:1. The average size of the wear particles was approximately 40µm. Particles in this size range

are characteristic of abnormal wear.

Figure 5: Silhouettes of particles from field samples produced by laser imaging system

4.5. Total Ferrous Particle Concentration Measurements

The total ferrous particle concentrations of the field samples were evaluated using the LaserNet 230, Ferrocheck

2000, and Optima 5300 ICP. As shown in Figure 6a, the ferrous particle concentrations measured via the LN 230

and Ferrocheck magnetometer systems were higher than the iron concentrations detected by the ICP. The mean

ferrous concentration for the LN 230, Ferrocheck, and ICP were 25, 18.5 and 2.6 ppm respectively (Fig. 6b). At a

95% confidence level, the difference between the mean ferrous particle concentration determinations in the

magnetometer-based instruments was not statistically significant. While the LN reported higher iron

concentrations than the Ferrocheck, the correlation coefficient was 91% as shown in Fig. 6c. A correlation

coefficient greater than 90% is indicative of a strong positive correlation between two measurements. ICP

indicated iron levels that were 5-times lower than the magnetometer systems. The lower iron concentrations

measured by ICP spectroscopy were due to the inefficient transport of ferrous particles to the plasma emission

source. Hence ICP analysis was non-quantitative, while the magnetometer methods demonstrated an ability to

quantify large ferrous particles, which is critical for trending machine health.

9

Figure 6: a) Comparison of the iron concentrations detected by ICP and magnetometer methods b) 95% confidence intervals of the mean iron concentrations, c) correlation between two magnetometer systems

4.6. Large Ferrous Particle Counting

While measuring ferrous particle concentrations is helpful for trending wear rates and machine health, the

presence of high iron concentrations alone is not necessarily an indication incipient machine failure because

normal wear can produce fine iron particles that accumulate in lubricants over time. Furthermore, the large

particles that result from the high-energy surface interactions associated with severe wear contribute surprisingly

little to the total concentration of iron in an oil sample. As shown in Eqn. 1, the mass of a solid spherical particle is

proportional to density (𝜌) and the cube of the particle radius (𝑟).

𝑚𝑎𝑠𝑠 𝑜𝑓 𝑎 𝑠𝑝ℎ𝑒𝑟𝑒 = 43⁄ 𝜋 ∙ 𝑟3 ∙ 𝜌 𝐸𝑞𝑛. 1

Rarely are wear particles spherical in shape. Sliding wear particles for instance, tend to be flat and elongated with

approximate dimensional ratios 10:3:1. [11] As shown in Eqn. 2, the mass of a solid elipical particle is proportional

to length (l), width (w), height (h) and density (𝜌).

𝑚𝑎𝑠𝑠 𝑜𝑓 𝑎𝑛 𝑒𝑙𝑖𝑝𝑠𝑒 = 43⁄ 𝜋 ∙ (𝑙/2 ∙ 𝑤/2 ∙ ℎ/2) ∙ 𝜌 𝐸𝑞𝑛. 2

Assuming the density of steel is 7.8g/cm3, an elliptical sliding wear particle that is 50µm in longest chord with

10:3:1 dimensional proportions weighs approximately 1.70x10-8 grams. The typical density of a hydraulic oil is .87

g/cm3 at 15.6°C and therefore 1 ml of oil weighs approximately 0.87 grams. Hence, a fluid containing one 50µm

ferrous sliding wear particle per ml will have a theoretical iron concentration of 1.7e-8/0.87 = 0.0195 ppm. Based

on this analysis, a concentration of fifty 50µm ferrous sliding wear particles per ml would be required to raise the

iron concentration of a hydraulic fluid by 1ppm. Thus, high counts of large wear particles may be present in oil

samples when the part per million concentration of iron is low.

10

LN 230 ferrous particle counts in the >25µm particle size range are compared to iron concentrations in

Figure 7. The number of >25µm ferrous particle counts is shown in an increasing order from left to right.

The corresponding concentration of ferrous particles is indicated by the adjacent bar (shown in red). The

concentration of ferrous particles in ppm does not correlate with the ferrous counts. While these results

seem counterintuitive, if one considers the fact that fifty 25µm particles are required to produce a 1 ppm

increase in iron concentration, it is understandable that the two measurements do not correlate.

Figure 7: Comparison of large ferrous particle counts and ferrous particle concentration

4.7. Interpretation of Results

While the counts and concentrations of ferrous particles do not correlate, they are both indicators of machine

condition. Two parameters were used for characterizing machine condition based upon ferrous counts and

concentrations: Percentage of Large Ferrous Particles (PLFP) and Ferrous Wear Severity Index. The PLFP is

determined from the ratio of large to total ferrous particle concentration in ppm as shown in Eqn. 3.

𝑃𝐿𝐹𝑃 =𝐿𝑎𝑟𝑔𝑒 [𝐹𝑒]𝑝𝑝𝑚

𝑇𝑜𝑡𝑎𝑙 [𝐹𝑒]𝑝𝑝𝑚⁄ ∙ 100 𝐸𝑞𝑛. 3

For instance, in sample ten, the ferrous particle count >25µm was half the total ferrous particle concentration.

Converting the >25µm counts to concentration and calculating the PLFP value via Eqn. 3 reveals that only 4% of the

ferrous particles in sample ten were large. The Ferrous Wear Severity Index (FWSI) is determined from the product

of the total and large ferrous particle concentrations as shown in Eqn. 4.

𝐹𝑊𝑆𝐼 = 𝐿𝑎𝑟𝑔𝑒 [𝐹𝑒] 𝑝𝑝𝑚 ∙ 𝑇𝑜𝑡𝑎𝑙 [𝐹𝑒]𝑝𝑝𝑚 𝐸𝑞𝑛. 4

Samples ten and thirty have similar total ferrous particle concentrations but sample thirty has a higher

concentration of large ferrous particles. The FWSI indices for these samples were 16.5 and 61.2 respectively.

Thus, two oils with similar iron concentrations differ significantly in terms of wear severity. These results

demonstrate the value of combining magnetic and particle imaging sensors in an integrated system.

11

5. Summary and Conclusions

Particle imaging, electromagnetic, ferrographic, and spectroscopic methods were used to evaluate oil samples

collected from hydraulic machines known to be generating abnormal wear particles. The wear debris in the

samples was characterized in terms of ferrous particle size and concentration. The following observations were

made:

1) The mean contamination level of the oil samples was ISO Code 24/22/17 which exceeds the RCL of a typical

hydraulic system. On average, of 25% of the large particles were non-metallic while 75% were metallic.

2) Ferrographic analysis revealed the presence of large particles that appeared to have been flattened or

reworked as a result of being squeezed through rolling contacts. Numerous small rubbing wear particles were

also observed.

3) The magnetometer systems were 5-times more sensitive to the presence of large iron particles than ICP

atomic emission spectroscopy.

4) The concentration of ferrous particles in the oil sample did not correlate with the number of large particles

present because large particles resulting from the high-energy surface interactions associated with severe

wear contribute very little to the total concentration of iron in an oil sample.

5) Based on an analysis of the relationship between the counts of large particles and the volume of a typical wear

particle, fifty 50µm ferrous sliding wear particles per ml would be required to raise the iron concentration of

an oil sample by 1ppm.

6) Combining a particle imaging system with high-sensitivity magnetometers in an integrated oil analysis

instrument facilitates the quantification of large and small ferrous wear particles.

7) The ratio of large to small ferrous wear particles and their concentration revealed the severity of wear

occurring in the hydraulic system.

Acknowledgments: This material is based upon work supported by the National Science Foundation under Grant

No. 1263346; Research Experiences for Undergraduates in Fluid Power.

6. References 1) Fitch J.C., “Control and Management of Contamination in Hydraulic Fluids,” in Totten, G. E. (2011). Handbook of hydraulic

fluid technology, 2nd Ed. CRC Press, 2011, pp. 219-255.

2) ISO 21018-1:2008 Hydraulic fluid power – Monitoring the level of particulate contamination of the fluid – Part 1: General

principles

3) ISO 4406:2017 Hydraulic fluid power – Fluids – Method for coding the level of contamination by solid particles

4) ISO 12669 Hydraulic Fluid Power – Method for determining the required cleanliness level (RCL) for a system

5) Costandy S. Saba, Wendell E. Rhine, and Kent J. Eisentraut, Efficiencies of sample introduction systems for the transport of

metallic particles in plasma emission and atomic absorption spectrometry, Analytical Chemistry 1981 53 (7), 1099-1103

6) Muir, D. and Howe, B., "In-Line Oil Debris Monitor (ODM) for the Advanced Tactical Fighter Engine," SAE Technical Paper

961308, 1996, https://doi.org/10.4271/961308.

7) Howe, B. and Muir, D. 'In-Line Oil Debris Monitor (ODM) for Helicopter Gearbox Condition Assessment,' AD-a 347 503,

Defense Technical Information Center, Ft. Belvoir, VA, 1998.

8) Miller, Jerry L., and Duka Kitaljevich. "In-line oil debris monitor for aircraft engine condition assessment." In Aerospace

Conference Proceedings, 2000 IEEE, vol. 6, pp. 49-56.

9) Henning, PF. Olson, EJ. Glaberson, J. and Barraclough, T., “Particle Counter and Classification System,” U.S. Patent

9,274,041. 03 January 2016

10) ASTM D7596 – Standard Test Method for Automatic Particle Counting and Particle Shape Classification of Oils Using a

Direct Imaging Integrated Tester

11) ASTM D7684 – Standard Guide for Microscopic Characterization of Particles from In-Service Lubricants

12) ASTM D7690 – Standard Practice for Microscopic Characterization of Particles from In-Service Lubricants by Analytical

Ferrography

12

13) Verdegan, B.M., NIST Traceable Particle Counter Calibration Standard Contamination Control Enters the Next Century. SAE

transactions, 1998, pp.170-176.

14) GAF Chemicals, “Method of Producing Finely Divided Metals,” U.S. Patent 2,597,701. 20 May 1952.